Applications of Large Models in Games
Published:
Large models are moving into games as vision-action agents, hierarchical planners, companions, NPCs, production copilots, and dynamic content generators.
Technical Blog
Notes on reinforcement learning, multimodal agents, and visual generation.
从 Flow-GRPO 到生成式智能体：视觉生成强化学习正在走向闭环交互
Published:
从 Flow Matching、Flow-GRPO 与 DiffusionNFT 出发，讨论视觉生成强化学习如何从优化单次采样轨迹走向多轮环境交互。